Hydraulic soil properties of peatlands treatingmunicipal wastewater and peat harvesting runoff

Hydraulic soil properties of peatlands treating municipal wastewater and peat harvesting runoff

Anna-Kaisa Ronkanen & Bjørn Kløve

Anna-Kaisa Ronkanen & Bjørn Kløve, Water Resources and Environmental Engineer- ing Laboratory, Department of Process and Environmental Engineering, P. O. Box 4300, FIN-90014 University of Oulu, Finland, e-mail: anna-kaisa.ronkanen@oulu.fi Peat hydraulic conductivity (K), specific yield (S), degree of humification and shear strength were measured at two wetland treatment systems constructed on natural peatlands receiving different wastewater quality and loading in Northern Finland. Peat K was measured with a falling head piezometer test in situ and by taking soil cores in horizontal and vertical directions using Eijkelkamp cylinders. Peat S was obtained from pF-curves and drainage tests. The K in situ was 5.2 ´ÿ10^–7 – 2.9 ´ÿ10^–3 m s^–1, the horizontal K was 6.1 ´ÿ10^–6 – 3.8 ´ÿ10^–2 m s^–1 and the vertical K was 4.2 ´ÿ10^–6 – 2.6 ´ 10^–2 m s^–1. The highest K value was usually found in the vertical direction. The esti- mated acrotelm layer with high K reached 40–60 cm at Kompsasuo wetland and 10–60 cm at Ruka wetland. There was an agreement between different measurement methods for S when pF values corresponding to relevant negative pressure were used. S varied from 0.023 to 0.23. After several years of wastewater loading, the peat hydraulic con- ductivity was still sufficient to maintain wastewater flow in the top 50 cm of the peatland.

Keywords: hydraulic conductivity; peat; specific yield; wastewater treatment; wetland

Introduction

In Finland, natural peatlands were used from 1957 to early 1980s to treat municipal wastewater by infiltrating the wastewater into ditches on peat- lands (Surakka & Kämppi 1971, Lehtonen 1994).

The first experiments showed that the high solid loads clogged the peat, and the method was there- fore not suitable in primary or secondary treat- ment with high solid loading (Munsterhjelm 1972, Lehtonen 1994). However, since the late 1980s, peatlands have been used to treat waters with rela- tively low suspended solid content such as peat harvesting runoff and forest drainage runoff (e.g.

Ihme et al. 1991, Heikkinen et al. 1994, Savolainen et al. 1996, Lyytikäinen et al. 2003). In 1994, a pilot project was started in Ruka (Finland) to com- plement chemically treated municipal wastewater by polishing the treated wastewater in peatland (Pirttijoki 1996, Hallikainen 2003). During the pilot projects, some design and management guidelines were formulated. However, the hydrau- lic properties such as peat hydraulic conductivity or specific yield had not been previously meas- ured, and flow depths of wastewater in peatland remained uncertain. The active flow layer is im- portant as it determined the total soil and water volume available for purification.

The soil physical properties of natural peatland are generally layered, with a top layer, acrotelm, consisting of living plants and rapidly decaying plant material and a lower layer, catotelm, where the plants have been humified to peat. In peatlands the hydraulic conductivity decreases with increasing humification and depth (Huikari 1959, Boelter 1969, Korpijaakko &

Radforth 1972, Päivänen 1973). It is well docu- mented that the hydraulic conductivity of catotelm is low at around 10^–7–10^–9 m s^–1 (Boelter 1965, Chason & Siegel 1986), whereas the hydraulic conductivity of acrotelm peat can be as high as 10^–1 m s^–1 (Burt et al. 1990 and Hobbs 1986).

There is usually a large spatial variation in the hydraulic properties of peat and the hydraulic conductivity depends to a large degree on the pore-size distribution of the peat. Peats with high fiber content and low bulk density have the high- est hydraulic conductivity (Boelter 1965). Also, the variation between mires can be high. Huikari (1959) found that the hydraulic conductivity was higher in the upper peat layer in bog fens than in pine mires, but the contrary was found in deeper peat layers (over 30 cm).

In addition to hydraulic conductivity, the sub- surface flow in peat depends on peat water stor- age capacity. The effective porosity affects the subsurface flow velocity and water table fluctua- tions. In general, the water table movement in unconfined systems such as the acrotelm is mod- elled using the concept of soil specific yield, which is the relationship between the quantity of water that has been added to or removed from the soil and the subsequent change in the water table (Boelter 1965). Previous studies indicate that the specific yield decreases from around 0.8 to 0.1 as the peat changes from an undecomposed peat to a well-decomposed peat (Boelter 1965, Päivänen 1973).

Objectives of the study

The aim of this study was to determine the hy- draulic properties of two peat-based wetland treat- ment systems which have been loaded with wastewater for several years. The hydraulic prop- erties are a prerequisite for estimating water resi- dence time and for modelling flow and transport

of nutrients in the wetlands. In previous wetland studies, it has been suggested that the accumula- tion of organic matter could reduce the perme- ability and the life-time of the wetland (Sundblad 1988, McIntyre & Riha 1991, Tanner & Sukias 1995, Tanner et al. 1998, Mæhlum 1998). An- other objective of this study was to determine the effective or main flow depth in which the wastewater flows in the peatland treatment sys- tems. This is important as the purification proc- esses occur mainly in the active flow layer.

Material and methods Study Sites

The sites are located in northern Finland in the municipalities of Kuivaniemi (Kompsasuo, 65º45’N, 26º00’E) and Kuusamo (Ruka, 66º10’N, 29º7’E) (Fig. 1). The sites selected had similar peat properties, but received different types of wastewater. Kompsasuo wetland purifies drain- age water from a peat harvesting area and Ruka wetland purifies municipal wastewater from a skiing resort after conventional chemical treat- ment. Both wetlands were originally minero- trophic mires with Sphagnum and Carex peat, varying from von Post humification scale (e.g.

Puustjärvi 1970) of H1 to H5 in the upper 1 m layer (Table 1). The wetland areas receiving wastewater are 2.4 ha and 0.8 ha at Kompsasuo and Ruka, respectively. The sites have been loaded with wastewater since 1987 and 1995 for Kompsasuo and Ruka, respectively.

Wastewater is fed continuously to wetlands by distribution ditches in upper parts of the wetlands (Fig. 1). Both horizontal subsurface and surface flow occurs in both wetlands. At Ruka, the wetland is by-passed if the wastewater load exceeds 40 m³ h^–1. A more detail description of Kompsasuo wetland is given in Ihme et al. (1991).

In frost-free period from June to October, the hydraulic load at Kompsasuo is on average 17 mm d^–1, but can reach up to 54 mm d^–1 during peak runoff (Table 1). The wetland removes on annual average 25–64% of a total nitrogen load, 20–64% of a total phosphorus load and 21–64%

of a suspended solids load in the frost-free pe- riod. At Ruka wetland, the hydraulic load is on

Table 1. Characteristics of the study sites on wetland (Kompsasuo and Ruka).

Taulukko 1. Kompsasuon ja Rukan tutkimuskosteikkojen ominaisuustietoja.

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Kompsasuo Ruka

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Area (ha) 2.4 0.8

Slope (‰) 8 8.5

Design criterion 4.8% of drainage area Max load 250 m³/d

Peat depth (m) 1.9–3.1 0.5–1.5

Peat type¹ SC and MenCS CS

Degree humif. (von Post) H1–H5 H1–H6

Aver. Min. Max. Aver. Min. Max

–––––––––––––––––––––––– ––––––––––––––––––––––––

Discharge to

the wetland mm d^–1 17 254² 36 6 75

N kg ha^–1 d^–1 1 0.007 19³ 2 1 1 79⁴

mg l^–1 20.3 7 41 3 81

Total P kg ha^–1 d^–1 0.01 0.001 0.3 0.20.01 0.7

mg l^–1 0.06 0.01 0.3 0.4 0.06 2

BOD₇ kg ha^–1 d^–1 0.1 0.05 0.23 0.214

mg l^–1 2 0.5 4 6 1 28

Susp. solids kg ha^–1 d^–1 1 0.04 35 6 0.3 25

mg l^–1 6 1 104 3 2 34

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

1 Men = Menyanthes, C = Carex, S = Sphagnum: ² Frost–free period (from June to October): ³ total N: ⁴ Inorg.N Fig. 1. Location of Kompsasuo (a) and Ruka (b) wetlands. (Photo: Suomen Ilmakuva Oy)

Kuva 1. Kompsasuon (a) ja Rukan (b) tutkimuskosteikkojen sijainti. (kuva: Suomen Ilmakuva Oy)

average 36 mm d^–1 reaching an annual maximum during winter ski holidays. The wetland reduces efficiently total phosphorus (69–93%), suspended solids (68–94%) and BOD (32–94%). The nitro- gen load is mainly inorganic nitrogen and it is removed about 5–43% in the wetland.

Sampling points and soil sampling

Soil samples were taken in the treatment wetlands and for comparison from the adjacent natural peatland not affected by wastewater. At Kompsasuo, samples were taken from 9 sampling points in the wetland and from 3 sampling points in the reference area (Fig. 2a). At Ruka, 7 sam- ples were taken in the wetland and one in the ref- erence area (Fig. 2b). During sampling, the wa- ter table in both wetlands was at 0–5 cm below soil surface.

Intact peat cores were taken from 65–85 cm depths using a sharp edge auger (8 cm x 8 cm).

For hydraulic conductivity measurements, 100 cm² cylinders (radius 2.5 cm) were pushed into the augered peat samples in the vertical and hori- zontal directions. No compression was observed.

Furthermore, pF-rings (46.6, 51.3 or 54.0 cm³) were taken from the augered peat for water re- tention measurements at Kompsasuo. At Ruka, undisturbed soil cylinders (8 cm in diameters) of 10, 20 and 30 cm height were taken for determi-

nation of soil water retention.

The shear strength was measured in situ with a hand vane tester (GeoNor), using 20 mm by 40 mm vane. At all sites, samples were also taken to determine the degree of peat decomposition (von Post), peat type and bulk density. Correlation analysis was used to test dependence of the hy- draulic conductivity on other physical peat prop- erties.

Specific yield and soil moisture retention The specific yield (S) was determined by a sim- ple drainage test and by calculation from soil water retention curves (pF-curves). The water retention curve for five different depths (4, 10, 30, 50 and 70 cm) was determined in a pressure cell for undisturbed fresh peat samples of Kompsasuo wetland. Wet peat samples were satu- rated and placed on a saturated ceramic plate. Peat water content was measured by weighing at eight pressure steps between pF 0 and 2.9. At the end of the experiment, peat samples were dried at 105ºC for 24 h. In between measurement points, the volumetric water content was obtained by lin- ear interpolation. S was calculated from water retention measurement for different assumed groundwater depths, i.e. negative pressure in peat.

The specific yield of Ruka wetland was cal- culated from the results of a drainage test similar

Fig. 2. Sampling points at a) Kompsasuo and b) Ruka wetlands and adjacent reference areas (R).

Kuva 2. Mittauspisteet a) Kompsasuolla ja b) Rukalla sekä niiden vertailualueilla (R).

to Tolman (1937). The drainage test was carried out using two 10 cm high, two 20 cm high and four 30 cm high intact peat cores that free drained on a wire tray for 13 days by gravity. The sam- ples were weighed after 0, 5 and 13 days. Evapo- ration was eliminated by covering the samples and keeping the samples in a cool and dark room.

The specific yield at corresponding pressure (10–

30 cm) was calculated as the ratio of water re- leased to original soil volume.

Hydraulic Conductivity

The saturated hydraulic conductivity was meas- ured in two different ways: with a time- consum-

ing laboratory method and a fast field method.

The peat K in situ was measured with a di- rect-push piezometer using the falling head method (Fig. 3). Due to head losses in the pi- ezometer the methods allows for accurate K- measurements below 0.2 cm s^–1.

The rate of the outflow (q) at the piezometer tip at any time (t) is proportional to the hydraulic conductivity (K) of the soil and to the unrecov- ered head difference (H–h) (Hvorslev 1951), so that

h) t FK(H

U h

q(t) ² = −

∂

= ∂ (1)

where r = radius of the piezometer reservoir;

H = water level at the measuring point; h = water level in the reservoir; F = shape factor; t = time.

Eq. (1) is rearranged in the following form

U H h FK U FK t h

2

2 =

∂ +

∂ (2)

which after integration, resolving the integra- tion constant from the initial condition h(0) = H0

and taking a natural logarithm yield U t

FK H

H H -

ln h ₂

0

−

⎟⎟⎠=

⎜⎜⎝ ⎞

⎛

− (3)

Eq. (3) represents a straight line on a semi- logarithmical graph when the left-hand side is plotted against time. The slope of the line gives K when F is known. The calibration of the F fac- tor was carried out with K measured in the labo- ratory (Eijkelkamp). Discharge of water (Q) through the porous media is defined by Darcy’s law written in a one-dimensional form:

qA xA K H

Q =

∂

= ∂ (4)

where A = the cross-sectional area of flow layer (m²), H = hydraulic head (m), x = distance in direction of flow (m) and q = specific discharge (m s^–1), then the flow velocity (v) is given by

Fig. 3. Schematic of the direct-push piezometer: L = length of the perforated outlet; R = radius of the piezometer.

Kuva 3. Kaaviokuva mittauksissa käytetystä pietsometris- tä: L = purkuaukon pituus; R = pietsometrin säde.

n

v=q (5)

where n = the effective porosity. The average detention time of water (td) can be estimated by

Q

t_d =V (6)

where V = water volume in a wetland (m³).

The detention time for each 10 cm layer was obtained using Darcy’s law, and the respective average K value to estimate Q and porosity to obtain V. The detention times were scaled with mean detention time for the entire 70 cm peat layer (t*).

Results

Degree of humification

The degree of humification varied from H1 to H5 (depth 0–80 cm) in Kompsasuo wetland. Typi- cally, the peat type was Sphagnum-Carex (H1–

H5) in the upper peat layers (0–30 cm) and Carex- Sphagnum (H4–H5) at 30–80 cm. No change in the degree of humification was found compared to measurements carried out in 1992 (Ihme 1994).

At Ruka the peat consisted of Carex-Sphagnum (H1–H5) in the upper part of the wetland (sam- pling points B3 and B2) and Sphagnum-Carex (H1–H5) in the lower part of the wetland (sam- pling point B1). At one measuring point (B1), the

degree of humification reached H6 already at the depth of 25–35 cm.

In general, the shear strength varied from 4 to 36 kPa at Kompsasuo and 8 to 41 kPa at Ruka wetland (Fig. 4) which is in agreement with Ca- nadian observation of 1.5 kPa–38 kPa for amor- phous or fine fibrous peat (Landva 1980), but somewhat lower than observed by Burke (1978).

At Kompsasuo, the shear strength systematically decreases as far as 30 cm from the surface, sup- porting Kløve’s (2000) observations, but then in- creased reaching generally the maximum value at a depth of 90–100 cm. At Ruka wetland, such a decrease was not observed even if maximum val- ues generally occurred at a depth of 90–100 cm.

Specific yield

Specific yield (S) for both sites varied from 0.023 to 0.11 at suction 10 cm H2O, from 0.045 to 0.14 at 20 cm H2O and from 0.068 to 0.23 at 30 cm H2O suction (Table 2 and 3). The lowest values were observed in the deepest peat layers and the highest values in the upper layers. All the peat sam- ples contained more than 82% water by volume when saturated. During the measurements, the volume of the samples slightly decreased due to drying. S values after five days in the drainage test coincide with specific yield determined from pF- curves for layers at depths 4–10 cm. At deeper lay- ers, S was some higher determined from the drain- age test. Furthermore, S after 13 days was slightly higher at depths 4–10 cm and clearly higher at deeper layers than S determined from pF-curves.

Table 2. The peat specific yield determined from pF–curves for Kompsasuo wetland.

Taulukko 2. Kompsasuon pF–käyrästä määritetyt ominaisantoisuudet.

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Specific yield

––––– 10 cm H₂O ––––– ––––– 20 cm H₂O ––––– ––––– 30 cm H₂O –––––

Depth (cm) Min. Ave. Max. Min. Ave. Max. Min. Ave. Max.

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

4 0.045 0.053 0.066 0.089 0.107 0.13 0.13 0.16 0.2

10 0.027 0.043 0.061 0.053 0.086 0.12 0.08 0.13 0.18

30 0.025 0.033 0.043 0.051 0.066 0.086 0.076 0.099 0.13

50 0.026 0.028 0.03 0.052 0.056 0.059 0.078 0.084 0.089

70 0.023 0.03 0.034 0.045 0.061 0.068 0.068 0.091 0.1

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Fig. 4. Shear strength in Kompsasuo and Ruka wetlands at the depths 0–100 cm below the peat surface.

Kuva 4. Kompsasuolla ja Rukalla mitatut leikkauslujuudet turpeen syvyyksillä 0–100 cm.

Table 3. The peat specific yield determined by the drain- age test for two drainage durations on samples from Ruka wetland.

Taulukko 3. Rukan valutuskokeessa määritetyt ominaisan- toisuudet kahdella eri valutusajalla.

––––––––––––––––––––––––––––––––––––––––––––

–––– Specific yield –––

Pressure (cm H₂O) 5 d 13 d

––––––––––––––––––––––––––––––––––––––––––––––

10 0.066 0.11

20 0.11 0.14

30 0.16 0.23

––––––––––––––––––––––––––––––––––––––––––––––

Hydraulic conductivity

At Kompsasuo wetland, the average in situ K was almost constant (1.0 ´ÿ10^–3 – 1.9 ´ÿ10^–3 m s^–1) for 0–50 cm depths (Fig. 5). At depths below 50 cm, a rapid decrease from 1.0 ´ÿ10^–3 cm s^–1 to 2.9 ´ 10^–6 m s^–1 was observed. In the top layer (0–10 cm), K varied between 1.0 ´ÿ10^–3 and 2.9 ´ÿ10^–3 m s^–1 whereas at a depth of 30–40 cm the K ranged from 5.6 ´ÿ10^–4 to 2.8 ´ÿ10^–3 m s^–1. The measure- ments indicate an increasing variation of K with

depth. At Ruka wetland, the average in situ K was nearly one order of magnitude smaller for depths of 10–40 cm than at Kompsasuo wetland (Fig. 5). The variation in K was high for all lay- ers and the variation did not increase with depth.

In the top layer, K ranged from 8.0 ´ÿ10^–5 to 2.4 ´ 10^–3 m s^–1. The mean value remained almost con- stant for depths 10–50 cm. Below 50 cm a rapid decrease was observed.

In the reference areas, the in situ K was gen- erally lower than in the treatment wetlands, vary- ing from 3.9 ´ÿ10^–6 to 9.1 ´ÿ10^–4 m s^–1 and from 2.3 ´ÿ10^–7 to 2.0 ´ÿ10^–4 m s^–1, for Kompsasuo and Ruka, respectively. The highest K values were observed at depths of 10–20 cm and the lowest values were observed at depths of 60–70 cm.

Generally, in these layers the average K was the same order of magnitude as in the treatment wetlands, but at all other depths, the K was one or two orders of magnitude lower in the refer- ence sites than in treatment wetlands. However, in Ruka wetland at depths 0–30 cm (sampling points B3 and B2), the minimum values were one or two orders of magnitude smaller than in the reference site.

The horizontal hydraulic conductivity (Kh) measured with Eijkelkamp cylinders varied from 1.8 ´ÿ10^–5 to 1.3 ´ÿ10^–2 m s^–1 for Kompsasuo and from 6.1 ´ÿ10^–6 to 3.8 ´ÿ10^–2 m s^–1 for Ruka (Ta- bles 4 and 5). The vertical hydraulic conductivi- ties (Kv) were slightly greater, ranging from 4.2

´ÿ10^–6 to 1.9 ´ÿ10^–2 m s^–1 for Kompsasuo and from 3.1 ´ÿ10^–5 to 3.2 ´ÿ10^–2 m s^–1 for Ruka.

At Kompsasuo, the highest correlation of K (in situ) was found with shear strength and depth (Table 6). As the shear strength values are not corrected for rod resistance, the values are only be used to evaluate the variation in K. The corre- lation with the specific yield was 0.37. The peat bulk density and the degree of humification did not correlate significantly with K. In Ruka, the best correlation was observed between K and the degree of humification (–0.54) (Table 6). The correlation coefficient was –0.67 and –0.40, for depth and shear strength, respectively. In all data, the highest correlation was founded between shear strength and K.

The vertical variation of K, porosity and the hydraulic gradient in the wastewater flow direc- tion was used to estimate the flow velocities of

Fig. 5. The in situ K (m s^–1) in Kompsasuo and Ruka wetlands at depths 0–70 cm below soil surface, n = number of measuring points.

Kuva 5. Kompsasuon ja Rukan kosteikkojen in situ K (cm s^–1) eri syvyyksillä, n = mittauspisteiden lukumäärä.

wastewater (Eq. 5) in different peat layers. The results show that the wastewater flows mainly at depths of 0–40 cm in Kompsasuo and at depths of 0–30 cm in Ruka treatment wetland (Fig. 6).

In both wetlands, the water detention time of the peat profile is dominated by layers 0–50 cm (t* = 0.40–1.2) (Fig. 6).

Discussion Shear strength

The increase of shear strength at depth is mainly caused by increased resistance of the vane rod.

Because the resistance of the vane rod was not taken into account, the deeper peat shear strength is probably overestimated. However, the meas- ured values can be used to compare peat struc- ture between measurement points. The shear measurements do not indicate a clear change in peat properties with wastewater application. If the wastewater application caused changes in the

peat, differences in shear strength between val- ues close to the inlet and values close to the out- let would be expected as the solid load and nutri- ent load is highest close to the inlet. At Kompsa- suo, slightly lower shear values were measured close to the inlet ditch but at Ruka the results do not indicate any changes in shear values between locations. At Kompsasuo, the high shear strength in the top layer of C measuring line probably re- sults from the high root density of shoots as ob- served by Kaasinen (2003).

Specific yield

Originally, the specific yield was determined by drying saturated material by gravity and calcu- lating the yield from the volume of drained water per unit of saturated bulk volume (Tolman 1937, Boelter 1965); but the drainage time has not been reported in previous studies. Boelter (1965) has determined specific yield to be 0.15 for herba- ceous, moderately decomposed peat and 0.10 for well-decomposed peat by using water-retention values obtained with 10 cm high undisturbed core samples. Corresponding values for moderately decomposed peat at Ruka was 0.07 after 5 day draining and 0.11 after 13 day draining. Similar increase of S with time has also been found in sandy soils (Dos Santos Junior & Youngs 1969).

The results indicate that the drainage time should be considered when specific yield is used in peatlands to estimate changes in the water ta- ble. Partly for that reason the S-term of peat will be overestimated in transient flow simulation if the pF-curve (that describes steady state situa- tion at certain negative pressure) is used.

Generally, the specific yields determined from pF-curves has been calculated at pF 2 which cor- responds to a matric suction of 100 cm H2O or a drainage depth of 1 m. At pF 2, the specific yield at Kompsasuo wetland varied from 0.21 to 0.48 which is typical for moderately decomposed peat in Finland (Päivänen 1973). These values are similar to Canadian measurements by Price (1992) who measured variations of specific yields from 0.1 to 0.5 with depth for Sphagnum in lysimeters.

The value of specific yield depends on the

Fig. 6. The average flow profile for Kompsasuo and Ruka wetlands. Sizes of arrows indicate the relative discharge at each depth. v = flow velocity, n = porosity and t* = scaled mean detention time

Kuva 6. Keskimääräinen virtausprofiili Kompsasuolla ja Rukalla. Nuolen koko kuvaa suhteellista virtaamaa eri sy- vyyksillä, v = virtausnopeus, n = huokoisuus ja t* = skaa- lattu keskimääräinen viipymä

pressure differences used in calculation. It is ob- vious that using values determined at pF 2, spe- cific yields obtained are too high for calculating water table fluctuation in peatlands where the water table fluctuations occur near the ground surface and at higher water content than at pF 2.

Results in Table 2 show that different S values should be applied depending on the groundwater fluctuation (pressure drop) if the water table movement is simulated using specific yield. For small water table fluctuations, such as a 10 cm fluctuation, the specific yield is around 0.02–0.07, whereas at higher water table fluctuations (30 cm) the specific yield is about 0.07–0.2. The results indicate a somewhat lower S value for peat than has been previously estimated.

Hydraulic conductivity

The results of K (in situ) measurements indicate that the catotelm starts at a depth of 40–60 cm for Kompsasuo and at 10–60 cm for Ruka. Observed in situ catotelm K values (5 ´ÿ10^–7 – 10^–4 m s^–1) are in agreement with previous Finnish observations of 10^–8 – 10^–4 m s^–1 (Päivänen 1973, Kløve 2000).

The in situ hydraulic conductivity for the acrotelm

(10^–4 – 3 ´ÿ10^–3 m s^–1) is somewhat higher than 10^–

3 m s^–1 previously observed by Chason & Siegel (1986), but somewhat lower than observed by Burt et al. (1990) and Hobbs (1986) (0.1 m s^–1).

The difference of K between treatment wetlands and reference sites is partly due to wa- ter table lowering as a consequence of ditching of treatment wetlands affecting also the reference sites. Ditching accelerates settling and decompo- sition in peat which reduces peat porosity. At the reference site, the structural change in peat could be faster than in the constructed wetland as the wetland is kept wet with high wastewater load- ing. The higher K in the treatment wetland could also be caused by increased erosion in peat caused by increased hydraulic loading. Internal erosion in peat and formation of underground channels has been noted on blanket peat, a peat type that forms in areas with high precipitation. These con- ditions could be comparable to conditions with high hydraulic loading; however, this needs more studies. Anikwe & Nwobodo (2002) have ob- served long-term (20 years) municipal waste dis- posal to increase hydraulic conductivity of soil (consisting of sand, silt and clay) in Nigeria. They concluded that it was due to erosion of the soil.

Chason & Siegel (1986) have reported that

Table 4. Hydraulic conductivities for Kompsasuo wetland, K_h = horizontal hydraulic conductivity, K_v = vertical hydraulic conductivity and K = the hydraulic conductivity in situ.

Taulukko 4. Hydraulinen johtavuus Kompsasuolla Kh = horisontaalinen hydraulinen johtavuus, Kv = vertikaalinen hyd- raulinen johtavuus ja K = in situ hydraulinen johtavuus.

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Depth K_h (m s^–1) K_v (m s^–1) K (m s^–1) log(K_h/K_v)

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

A1 5 cm 4.2 ´ 10^–3 1.8 ´ 10^–2 - –0.6

20 cm 3.6 ´ 10^–5 1.6 ´ 10^–3 1.0 ´ 10^–3 –1.6

40 cm 8.8 ´ 10^–4 4.2 ´ 10^–6 2.0 ´ 10^–3 2.3

B1 5 cm 1.3 ´ 10^–2 9.2 ´ 10^–3 - 0.2

20 cm 8.7 ´ 10^–4 9.4 ´ 10^–3 2.6 ´ 10^–3 –1

40 cm 1.1 ´ 10^–3 - 2.8 ´ 10^–3 -

C1 5 cm 1.2 ´ 10^–3 1.6 ´ 10^–2 - –1.1

20 cm 5.6 ´ 10^–4 1.5 ´ 10^–2 1.6 ´ 10^–3 –1.4

40 cm 1.8 ´ 10^–5 1.9 ´ 10^–2 1.8 ´ 10^–3 –3

R1 5 cm 4.7 ´ 10^–5 8.3 ´ 10^–4 - –1.2

20 cm 8.1 ´ 10^–6 2.5 ´ 10^–6 2.6 ´ 10^–3 0.5

40 cm 1.7 ´ 10^–6 4.4 ´ 10^–5 4.2 ´ 10^–4 –1.4

Mean 1.8 ´ 10^–3 8.1 ´ 10^–3 1.9 ´ 10^–3

Standard deviation 0.004 0.008 0.8 ´ 10^–3

Range 1.3 ´ 10^–2 1.9 ´ 10^–2 2.4 ´ 10^–3

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Table 5. Hydraulic conductivities for Ruka wetland, Kh = horizontal hydraulic conductivity, Kv = vertical hydraulic con- ductivity and K = the hydraulic conductivity in situ.

Taulukko 5. Hydraulinen johtavuus Rukalla K_h = horisontaalinen hydraulinen johtavuus, K_v = vertikaalinen hydraulinen johtavuus ja K = in situ hydraulinen johtavuus.

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Depth K_h (m s^–1) K_v (m s^–1) K (m s^–1) log(K_h/K_v)

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

B3 5 cm 2.3 ´ 10^–2 1.2 ´ 10^–2 – 0.3

20 cm 2.3 ´ 10^–2 2.1 ´ 10^–2 4.2 ´ 10^–6 0

40 cm 5.8 ´ 10^–5 5.3 ´ 10^–5 2.9 ´ 10^–6 0

B25 cm 7.9 ´ 10^–4 2.6 ´ 10^–2 – –1.5

20 cm 2.9 ´ 10^–4 1.7 ´ 10^–2 3.1 ´ 10^–5 –1.8

40 cm 6.1 ´ 10^–6 3.1 ´ 10^–5 1.7 ´ 10^–5 –0.7

B1 5 cm 7.2 ´ 10^–3 2.2 ´ 10^–2 – –0.5

20 cm 3.8 ´ 10^–2 1.3 ´ 10^–2 1.8 ´ 10^–4 0.5

40 cm 2.5 ´ 10^–5 3.2 ´ 10^–2 6.7 ´ 10^–6 –3.1

R 5 cm 6.7 ´ 10^–3 4.2 ´ 10^–3 – 0.2

20 cm 1.5 ´ 10^–5 4.0 ´ 10^–4 2.0 ´ 10^–4 –1.4

40 cm 7.9 ´ 10^–4 4.2 ´ 10^–3 2.0 ´ 10^–6 –0.7

Mean 8.3 ´ 10^–3 1.3 ´ 10^–2 5.5 ´ 10^–5

Standard deviation 0.01 0.01 0.8 ´ 10^–4

Range 3.8 ´ 10^–2 3.2 ´ 10^–2 2.0 ´ 10^–4

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Kh was generally one or two orders of magnitude greater than Kv for partially decomposed peat in a spring fen-raised bog complex in the Lost River Peatland. Also others have found Kh to be greater than Kv (Korpijaakko & Radforth 1972, Beckwith et al. 2003). However, in this study Kv was greater than Kh in 65% of samples and smaller only in 26% of samples. Furthermore Kv was nearly equal to Kh in 9% of the samples.

In general, Kv was one or two orders of mag- nitude greater than Kh, The anisotropy factor log(Kh/Kv), as according to Chason and Siegel (1986), deviate from zero more than 0.5 nearly for all samples indicating that the peat structure orientation could be predominantly vertical and the peat body is anisotropic at depths of 0–40 cm.

Chason & Siegel (1986) have explained that ver- tical orientation of living stems of Sphagnum cre- ates vertical water passageways between the stems, whereas pronounced stratification and horizontal planar passageways are formed when the plants died and fall over.

In Ruka wetland, the horizontal K was the highest being similar to vertical K in the measur- ing point closest to the wastewater discharge area (B3 in Fig. 2b). This was also observed in Kompsasuo. If the horizontal direction is domi-

nant which often is the case in peat soils, this observation indicates some erosion rather than some clogging be caused by high hydraulic load- ing which is the highest near the inlet section.

However, differences between Kh and Kv for a given depth partly reflect small-scale heteroge- neity within the wetland as Kh and Kv were meas- ured from separate but adjacent soil cores.

In Kompsasuo, the measured in situ K values were similar to the corresponding laboratory val- ues, whereas for Ruka the in situ K was one or- der of magnitude lower than the laboratory val- ues. For the entire K-dataset, in situ K was lower than the laboratory values in six samples, between the horizontal and vertical hydraulic conductivi- ties in six samples and greater than values in the laboratory in three samples (Tables 4 and 5). A larger variation was observed in the laboratory values (standard deviation = 0.004–0.01) than for the in situ measurements (standard deviation = 0.0008–0.001) (Tables 4 and 5). Boelter (1965) and Päivänen (1973) have previously reported laboratory hydraulic conductivities to be greater than hydraulic conductivities in the in situ method. Probable reasons for differences are: i) the peat sample for the laboratory method repre- sents a very small part of the peat column, ii) there

could be leakage along the interface between the core and the inside wall of the sampling cylinder during the measurements, especially with peat materials having very low hydraulic conductiv- ity and iii) taking peat samples with the cylinders and transferring them from the field to the labo- ratory is difficult without breaking the peat struc- ture. Also, gas bubble formation during storage has been found to change the hydraulic conduc- tivity of the peat (Beckwith & Baird, 2001), but then the laboratory values would be somewhat lower than measured in the field.

The study sites are relatively similar in their degree of humification, peat type and shear strength, so it could be assumed that variation of hydraulic conductivities are also similar. How- ever, in situ K for the measurement point closest to wastewater discharge points (B3 and B2 in Fig.

2) for depths of 10–40 cm in Ruka wetland, were two or three orders of magnitude lower than hy- draulic conductivities for other points at equal depths. Furthermore, the lowest values in the top layer were in the observed main flow path of wastewater in Ruka. The results of hydraulic con- ductivity measurements might indicate some clogging in the main wastewater flow path in Ruka wetland. The average suspended matter loading-rate at Ruka was higher (5.6 kg ha^–1 d^–1) than at Kompsasuo (1.2 kg ha^–1 d^–1) which could be the reason why the results seem not to support clogging of peat at Kompsasuo.

The calculations indicate that the effective flow depth, where purification occurs in these peatlands, can be estimated to be around 50 cm. Hydraulic conductivities of these layers varied from 10^–3 to 10^–4 m s^–1. The effective flow depth observed in

this study is higher than the 0.20 m previously es- timated by Ihme (1994) indicating that a larger peat volume can be used for purification.

At the depth of 50–60 cm purification proc- esses could be achievable in the point of flow condition but below 60 cm the water flow is slow and in the same range as the diffusion rate. Hoag

& Price (1995) have found diffusion to be an important transport mechanism as the hydraulic conductivity of the peat was 10^–7–10^–6 m s^–1. Simi- lar values can be found at depths of 20–70 cm in Ruka and at depths of 50–70 cm in Kompsasuo, but generally below 60 cm in both wetlands.

Conclusions

The hydraulic conductivity for acrotelm varied from 3 ´ÿ10^–3 to 10^–4 m s^–1 and for catotelm from 5 ´ÿ10^–7 to 10^–4 m s^–1. Generally, the vertical con- ductivity was higher than the horizontal conduc- tivity. The laboratory method seemed to overes- timate the hydraulic conductivity compared to the in situ measurement.

The specific yield varied from 0.023 to 0.23, depending of the negative pressure used in cal- culating S values. The S value depended on time, and higher S values were obtained with long drainage times indicating that this should be ac- counted for if the S terms are used to explain vari- ation in water table fluctuations. There was an agreement between the drainage test and the pF- curve method determining the specific yield.

The study indicates that fairly high hydraulic conductivity is maintained in the peatlands de- spite several years of wastewater loading. The

Table 6. Correlations of the hydraulic conductivity between other physical properties, n = number of samples.

Taulukko 6. Hydraulisen johtavuuden korrelaatio turpeen muiden fysikaalisten ominaisuuksien kanssa, n = näytteiden lukumäärä.

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Shear n Depth n Humif. n Bulk n Specific n

strength density yield

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Ruka –0.40** 51 –0.67** 52–0.54** 25 - - - -

Kompsasuo –0.67** 78 –0.70** 79 –0.16 76 0.23 25 0.37* 25

All data –0.59** 129 –0.67** 131 –0.31** 101 0.23 25 0.37* 25

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

** significant at the 0.01 level and * at the 0.05 level

effective flow depth for both sites was estimated at approximately 50 cm, which is deeper than previously estimated.

Acnowledgements

The material of this study has been gathered and analysed in the Department of Biophysics in the University of Oulu. The authors would like to ac- knowledge the financial support of this project from the European Commission (PRIMROSE contract number EVK1-CT-2000-00065), Acad- emy of Finland, North Ostrobothnia Regional En- vironment Centre (NOREC), VAPO Oy, Energy and Water Cooperative of Kuusamo (EVO) and the city of Kuusamo. The manuscript was re- viewed by Jan Vymazal and Niko Silvan.

References

Anikwe, M. & Nwobodo, K. 2002. Long-term effect of municipal waste disposal on soil properties and pro- ductivity of site used for urban agriculture in Abaka- liki, Nigeria. Bioresource Technology 83: 241–250.

Beckwith, C.W. & Baird, A.J. 2001. Effect of biogenic gas bubbles on water flow in poorly decomposed blanket peat. Water Resources Research 37: 551–558.

Beckwith, C.W., Baird, A.J. & Heatwaite. A.L. 2003. Ani- sotropy and depth-related heterogeneity of hydraulic conductivity in a bog peat. I: laboratory measurements.

Hydrological Processes 17: 89–101.

Boelter, D.H. 1965. Hydraulic conductivity of peat. Soil Science 100: 227–231.

Boelter, D.H. 1969. Physical properties of peat as a related to degree of decomposition. Division S-6 – Soil and Water Management and Conservation, pp. 606–609.

Burke, W. 1978. Long-term Effects of Drainage and Land Use on Some Physical Properties of Blanket Peat. Irish Journal of Agricultural Research 17: 315–322.

Burt, T.B., Heathwaite, A.L. & Labadz, J.L. 1990. Runoff production in peat covered catchments. In: M.G. An- dersson and T.P. Burt (eds), Process studies in hills- lope hydrology, John Wiley, pp. 463–499.

Chason, D.B. & Siegel, D.I. 1986. Hydraulic conductivity and related physical properties of peat, Lost River Peat- land, Northern Minnesota. Soil Science 142: 91–99.

Dos Santos Junior, A.G. & Youngs, E.G. 1969. A Study of the specific yield in land-drainage situation. Journal of Hydrology 8: 59–81.

Hallikainen, S. 2003. Kosteikkokäsittely Rukan ja Kempe- leen jälkikäsittelykentillä: puhdistustehon vaihtelu ja kehitys (Summary: Wetland treatment in wastewater

plants of Ruka and Lakeus: Changes in reduction rates and development of efficiency). M.Sc. Thesis (Tech.), University of Oulu, 99 pp.

Heikkinen, K., Ihme, R. & Lakso, E. 1994. Ravinteiden, orgaanisten aineiden ja raudan pidättymiseen johtavat prosessit pintavalutuskentällä (Summary: Processes contributing to the retention of nutrients, organic mat- ter and iron in an overland flow wetland treatment sys- tem). Publications of the Water and Environment A193, National Board of Water and Environment, Helsinki, Finland, 80 pp.

Hoag, R.S. & Price, J.S. 1995. A field-scale, natural gradi- ent solute transport experiment in peat at a Newfound- land blanket bog. Journal of Hydrology 172: 171–184.

Hobbs, N.B. 1986. Mire morphology and the properties and behaviour of some British and foreign peats. Quar- terly Journal of Engineering Geology 19: 7–80.

Huikari, O. 1959. Kenttämittaustuloksia turpeiden veden- läpäisevyydestä (Referat: Feldmessungsergebnisse über die wasserdurchlässigkeit von torfen). Communica- tiones Instituti Forestalis Fenniae 51:1–26 pp.

Hvorslev, M.J. 1951. Time lag and soil permeability in ground-water observations. Waterways Experiment Station, Corps of Engineers, Bulletin no 36, 50 pp.

Ihme, R. 1994. Pintavalutus turvetuotantoalueiden va- lumavesien puhdistuksessa (Summary: The use of over- land flow for the purification of runoff water from peat mining areas). Ph.D.Thesis, University of Oulu, 140 pp.

Ihme, R., Heikkinen, K. & Lakso, E. 1991. Pintavalutus turvetuotantoalueiden valumavesien puhdistuksessa (Summary: The use of overland flow for the purifica- tion of runoff water from peat production areas). Pub- lications of the water and environment research insti- tute, National Board of Waters and Environment, Hel- sinki, Finland A75, 181 pp.

Kaasinen, E. 2003. Kompsasuon pintavalutuskentän kas- villisuusmuutoksista, M.Sc. Thesis, University of Oulu, 42 pp.

Kløve, B. 2000. Effect of peat harvesting on peat hydrau- lics properties and runoff generation. Suo 51: 121–129.

Korpijaakko, M & Radforth, N.W. 1972. Studies on the hydraulic conductivity of peat. The Proceedings of the International Peat Congress, Volume III, pp. 323–334.

Landva, A.O. 1980. Vane testing in peat. Canadian Geo- technical Journal 17: 2–19.

Lehtonen, J. 1994. Jäteveden kehitys Suomessa pitkällä aikavälillä (Summary: Long-term development of waste water treatment in Finland). Tampere University of Technology, Institute of Water and Environmental En- gineering No. B 58 100 pp.

Lyytikäinen, V., Vuori, K-M. & Kotanen, J. 2003. Kunnos- tusojitusten suojavyöhykkeiden toimivuus – Kuohatti- järven pintavalutuskenttien tutkimukset vuosina 1998–

2001 (Summary: Functioning of the buffer zones in forest drainage sites – results of the overland flow area studies in the Kuohattijärvi catchment in 1998–2001), Regional Environmental Publications 315, North Kare- lia Regional Environment Centre, Joensuu, 62 pp.

McIntyre, B.D. & Riha, S.J. 1991. Hydraulic conductivity and nitrogen removal in artificial wetland system. Jour- nal of Environmental Quality 20: 259–263.

Munsterhjelm, K. 1972. Jäteveden suoimeyttämöiden toiminnasta eri vuodenaikoina ja sen tehostamiseta (Summary: On the efficiency of peat infiltration of sew- age during various seasons and on its improvement).

National Board of Waters, Finland, Report 38, 114 pp.

Mæhlum, T. 1998. Cold-climate constructed wetland: Aer- obic pre-treatment and horizontal subsurface flow wet- lands for domestic sewage and landfill leachate Purifi- cation. Doctoral Thesis, Agricultural University of Norway, 180 pp.

Pirttijoki, J. 1996. Pintavalutus Kuusamon Rukan jäteves- ien jälkikäsittelymenetelmänä (Summary: Treatment of wastewater from Ruka in Kuusamo by using overland flow method). M. Sc. Thesis (Tech), University of Oulu, 85 pp.

Price, J.S. 1992. Blanket bog in Newfounland. Part 2. Hy- drological processes. Journal of Hydrology 135: 103–

119.

Tiivistelmä: Turpeen hydrauliset ominaisuudet kunnallisten jätevesien ja turvetuotanto- alueen valumavesien puhdistukseen rakennetuissa kosteikkopuhdistamoissa

Tässä tutkimuksessa mitattiin turpeen hydraulista johtavuutta (K), ominaisantoisuutta (S), maatunei- suusastetta ja leikkauslujuutta kahdella luonnontilaiselle suolle rakennetulla jätevesien kosteikko- puhdistamolla Pohjois-Suomessa. Kosteikot ovat erilaiset puhdistettavan veden laadun ja määrän suhteen. K mitattiin muuttuvapaineisella pietsometrillä kentällä sekä Eijkelkampin sylintereillä labo- ratoriossa (sekä horisontaalinen että vertikaalinen K). Turpeen ominaisantoisuus määritettiin pF-käy- rästä sekä valutuskokeella. K in situ oli 5.2 ´ÿ10^–7 – 2.9 ´ÿ10^–3 m s^–1, horisontaalinen K 6.1 ´ÿ10^–6 – 3.8

´ÿ10^–2 m s^–1 ja vertikaalinen K 4.2 ´ÿ10^–6 – 2.6 ´ÿ10^–2 m s^–1. Vertikaalinen K oli tavallisesti suurempi kuin horisontaalinen K. Arvioitu virtauskerroksen paksuus oli 40–60 cm Kompsasuolla ja 10–60 cm Rukalla. Eri ominaisantoisuuden (S) mittausmenetelmien välillä oli hyvä yhteensopivuus. Ominais- antoisuus vaihteli välillä 0.023–0.23. Usean vuoden kuormituksesta huolimatta tutkimuskosteikko- jen hydraulinen johtavuus oli riittävän korkea mahdollistamaan jäteveden virtauksen noin 50 cm syvyisessä kerroksessa.

Puustjärvi, V. 1970. Degree of decomposition. Peat & Plant News 4: 48–52.

Päivänen, J. 1973. Hydraulic conductivity and water reten- tion in peat soils. Acta Forestalia Fennica 129: 1–70.

Savolainen, M., Kaasinen, A., Heikkinen, K., Ihme, R., Kämä, T. & Alasaarela, E. 1996. Turvetuotannon vesiensuojeluvaihtoehtojen tapauskohtainen vertailu, Suomen ympäristö 35. 116 pp.

Sundblad, K. 1988. Recycling of waster nutrients in a wet- land filter. Linköping Studies in Arts and Science No.

Surakka, S. & Kämppi, A. 1971. Jätevesien suoimey-27.

tyksestä. Suo 22: 51–57.

Tanner, C.C. & Sukias, J.P. 1995. Accumulation of organic solids in gravel-bed constructed wetlands. Water Sci- ence and Technology 32: 229–239.

Tanner, C.C., Sukias, J.P. & Upsdell, M.P. 1998. Organic matter accumulation during maturation of gravel-bed constructed wetlands treating farm dairy wastewater.

Water Research 32: 3046–3054.

Tolman, C.F. 1937. Ground water. McGraw-Hill Book Company, Inc., New York and London, 593 pp.

Received 30.9.2004, Accepted 4.2.2005